This article was written by

Follow

NEW YORK — The New York Times Connections Sports Edition puzzle No. 605 for Thursday, May 21, 2026, featured 16 sports-related words grouped into four categories centered on college game elements, pitching mound synonyms, Alpine skiing disciplines and Charlotte-based teams.

The puzzle, available through The Athletic, challenged players to sort the words into four groups of four. It carried a difficulty rating consistent with recent editions in the sports-themed variant.

The yellow category, typically the easiest, was “Seen at a college sporting event”: BAND, CHEERLEADERS, DANCE TEAM, STUDENT SECTION.

The green category was “Pitching mound”: BUMP, HILL, MOUNT, RUBBER.

The blue category was “Alpine skiing disciplines”: COMBINED, DOWNHILL, SLALOM, SUPER-G.

The purple category, often the hardest, was “Charlotte ____”: 49ERS, FC, HORNETS, NORTH.

Connections Sports Edition, a collaboration between The New York Times and The Athletic, adapts the original Connections format with a sports focus. Players identify connections drawing from athletics, teams, rules and sports culture. The game resets daily at midnight local time and allows four mistakes.

Hints for puzzle 605 included “Spotted at a game” for the yellow group, “Charlie Brown’s baseball home” for the green group, “Snow sport” for the blue group and “Queen City teams” for the purple group.

Yesterday’s puzzle on May 20, No. 604, included categories such as actions on a field goal attempt, U.S. Olympic host cities, Triple-A baseball teams and words ending in an NFL quarterback’s name.

The sports edition launched from beta and remains accessible via The Athletic’s platform or app, separate from the main NYT Games app. It targets fans with knowledge spanning multiple sports.

Players shared solve grids on social media, with many noting the college atmosphere group as accessible due to familiar game-day elements. The skiing disciplines required specific winter sports knowledge, while the Charlotte reference drew on NFL, MLS, NBA and college teams.

The pitching mound synonyms tested baseball terminology familiarity, with “rubber” as the official term for the pitcher’s plate. Alpine skiing categories highlighted Olympic and World Cup events.

Community discussions highlighted the puzzle’s mix of mainstream and niche sports references. International players sometimes found U.S.-centric elements like college game traditions more challenging.

The New York Times and The Athletic publish companion content with progressive hints. For May 21, hints guided solvers toward game-day visuals, baseball infrastructure, snow sports and regional team nicknames.

Connections Sports Edition builds on the popularity of the original game while incorporating trivia from football, baseball, basketball, skiing and other disciplines. No word repeats across puzzles, with editors curating balanced difficulty.

Strategies include scanning for obvious clusters like team names or event-specific terms. Players often start with yellow groups before tackling purple. Non-theme words serve as distractors to increase challenge.

Tomorrow’s puzzle, No. 606 for May 22, will present a new grid at midnight. The Athletic offers a Coach section for additional hints and player discussions.

The game supports engagement across devices through web and app access. Features include color-coded feedback and shareable results using emoji grids.

As of midday May 21, online forums discussed solve paths for puzzle 605, with users comparing times and noting near-misses involving overlapping sports terms. Many completed it within the allowed mistakes after identifying the college and skiing groups.

Educational value includes reinforcing sports knowledge, geography of host cities and terminology across leagues. The puzzle format promotes logical grouping and elimination strategies.

Connections Sports Edition continues to grow as part of The Athletic’s puzzle offerings alongside the main Connections. Puzzle 605 exemplified the blend of accessible and specialized categories that defines the series.

Q4: 2026-05-19 Earnings Summary

EPS of $0.30 beats by $0.00

 | Revenue of $1.40B (44.51% Y/Y) misses by $8.91M

This article was written by

Follow

On 06 April, a false tweet about President Donald Trump’s plans for tariffs led the S&P 500 to swing eight percentage points in minutes.

However, the $2.7 trillion market rebound wasn’t driven by traders themselves, but by algorithmic trading bots trained to react within seconds to announcements that might affect the stock market.

Speaking to Bloomberg, Benn Eifert Managing partner and co-chief investment officer at the hedge fund QVR Advisors said “You have a daisy chain of buying reactions in response to a headline”. He added, “algorithms are tuned to react extremely quickly to any kind of headline reversing tariffs.”

It was a reminder, if one was needed, that AI is already influencing markets.

AI is already dominating algorithmic trading

It is estimated that 70-80% of trading volume on the US stock market is executed through AI and algorithmic trading systems. Globally, the figures are even higher. Almost 89% of global trading volume is handled by AI-driven algorithms.

As AI tools have become more powerful, hedge funds and algorithmic traders have leaned increasingly on the technology; and for good reason too.

A few years ago, you would have had good reasons to doubt AI’s precision. But now, there can be no questioning its ability to process huge volumes of data in real-time and make close-to accurate forecasts.

How has AI changed algorithmic trading?

Think of AI as a multiplier of an algorithm’s ability. Satellite imagery of retail car parks, earnings call transcripts, social media sentiment, breaking news: AI can ingest all of it in real time and extract signals from sources that would have been invisible, or at least opaque a decade ago, even to the most advanced algorithm.

AI helps traditional algorithmic trading models move beyond constrained rule-based systems. AI models can learn and adapt, improving their efficiency and trades. The level of precision this produces marks a significant evolution from the rule-based algorithmic trading systems of the previous generation.

Humans: the often-forgotten driver of AI, and algorithmic trading

One element that is frequently omitted in the AI trading discussion, and arguably the most important, is the human.

Rotem Farkash, AI and trading expert weighed in on the subject, “AI models are only as good as the conditions and data they were trained on.”

Farkash argued, “Automated systems enable firms to operate far beyond traditional human capabilities; that is a fact. However, human judgment, and input, remains vital to AI and algorithmic trading, at least for the present.”

24/7 markets and operational edge

AI and algorithmic trading models have also changed the timeframe of trading. Your AI model or algorithm can place trades 24/5, or longer. Gone are the traditional 6.5-hour trading days in one geography or another.

AI also enables ‘round the clock’ support and market monitoring, increased liquidity and the ability to respond swiftly to price movements and global events.

AI Trading Risks: flash crashes

As seen with the huge swing in the S&P 500 in early April, AI’s introduction to trading is not without risks.

The Warsaw Stock Exchange (WSE) also suspended trading this month for over an hour after a surge of automated high-frequency orders triggered a feedback loop of bot-driven sell orders the exchange’s own systems could not contain.

Warsaw’s WIG20 index had plunged 7% before the plug was pulled. It took manual human intervention to restore order and the WSE subsequently vowed to review its algorithmic trading regulations entirely.

Regulators have responded more broadly too. In February 2026, ESMA issued a supervisory briefing identifying seven areas of increased oversight focus, with AI in trading featuring among them.

AI is here to stay in Finance and Trading

There is no doubt that AI is here to stay in trading and the case for its use is formidable. AI can extract signals from data that would have taken human analysts weeks to process. But it is not risk free.

For now, AI is only as good as the data and conditions it has been built on. What it looks like when that changes will be one the most consequential open questions in finance.